Article: A Midsummer Night’s Dream - Compare Simultaneous DSD64 and DSD256 Session Recordings

[vinoboy]

Just seeing this.. Have HAPI with AD8P/DA8P cards. Will test it out this evening

[wchazstory]

First time poster here. Downloaded the original files, using PC Laptop, JRiver 19, set to double DSD conversion, Geek out 720 analog into Marantz AV-8003, Rotel RMB-100 and Martin Logan Vista's. Limited to DSD 128, as the GO can't do 256.

 

As most people have stated I found the Horus files sounded better to me (and my wife). We both felt that the Horus DSD64 sounded better than the Grimm at DSD64. There was more weight, more air, instruments seemed to be more three dimensional in the sound field. In other words, the Horus files sounded more "real" for lack of a better word.

 

We struggled to hear much of a dfference between the Horus DSD64 and DSD128 tracks. I think I marginally heard a slight improvement in the DSED128 files, but I'm sure I would be hard pressed to pick which file is which if I were blindfolded. However, I think we could both identify the differences we heard between the Grimm DSD64 and Horus DSD64 files.

 

It was great to have a chance to hear the files and kudos to Channel Classics for making them Avaliable.

 

Chuck S.

[Jonalogic]

My first post, folks… be gentle, please!

 

Many thanks to Ted and Tom for the opportunity to listen to these tracks and beat up some of my preconceptions. As a scientist, I always welcome the chance to trash my ideas and have a re-think.

 

Herewith my feedback on these excellent tracks with Grimm 64, Horus 64 and Horus 128 recordings of the same excerpt from ‘A Midsummer Night’s Dream’. My Auralic Vega DAC meant I couldn’t evaluate the Horus 256, though.

 

The Mendelssohn is a piece that’s relatively easy to reproduce well, with good space, feather-light orchestration and nothing above mezzo forte.

 

Playback equipment was my normal system for hi-res digital downloads: An Auralic Aries bit cruncher, fed data from a high-speed Samsung 1Tb SSD; Auralic Vega DAC; Goldmund Mimesis 9 power amp and Martin Logan CLSIIz electrostatic speakers. Cabling in digital domain was by CAT, and Siltech monocrystal silver cable in analogue domain.

 

Anyway, my expectation was something like this: 1) that the Horus and Grimm would basically offer a different flavour but would not be radically different in overall SQ and 2) that the difference between the Horus 64 and Horus 128 would be quite subtle.

 

I was wrong on both counts. This was what I actually heard:

 

1) The Grimm 64 sounded markedly superior (IMO) to the Horus 64 – better focus, air, stage depth and timbral accuracy. Instruments sounded like the real thing.

2) The Horus 64 sounded more smeared in focus, less transparent, flatter in perspective and less musically accurate. I would hazard a guess that its time domain performance is/was not as good as that of the Grimm.

3) Horus 128 lifted the veil somewhat in focus and resolution, and also opened up some of the air and space. The difference between Horus 64 and 128 was more marked than I had expected.

4) However, I still preferred the Grimm overall. I’m not a Grimm groupie, but I have to say it impressed me mightily in these tests – mostly for musical reasons. Live music is my reference, and it simply sounded more to my ears like that!

 

And then Ted and Tom had to go and add two further tracks– respectively, the Horus DSD 256 downsampled to DSD 64 with Signalyst, and then the Signalyst-processed Grimm 64, with 0.34dB level tweak.

 

And here is where things started to get trickier. It would have been pretty easy to keep listening until I heard something but – in truth - comparing the two new tracks with the straight Grimm 64 these three seem to me, with my ears and playback equipment, now to be very much in the same ball-park, sonically. Yes, there are differences, but these seem much smaller than the large discrepancies that were pretty obvious between the straight Grimm 64 and downsampled Horus 64 and 128 tracks.

 

If this is valid, it suggests that the artefacts consequent on downsampling in the Merging DSD Converter software are sonically significant, whereas the Signalyst appears more transparent.

 

I would be happy musically with the straight Grimm or Signalyst tracks, less so it seems with those downsampled using Merging’s own software.

 

I hope this all makes sense!

[ted_b]

Jon, welcome to CA!! And thanks for the post. Yes, we are all becoming quite impressed with Miska's sampling routines. HQplayer is becoming a favorite, as I used it in my soon-to-be-finished exaSound e12 review, and now will use it to hopefully impress Jared with my surround setup (exaSound e28).

[chrille]

My first post, folks… be gentle, please!

 

Many thanks to Ted and Tom for the opportunity to listen to these tracks and beat up some of my preconceptions. As a scientist, I always welcome the chance to trash my ideas and have a re-think.

 

Herewith my feedback on these excellent tracks with Grimm 64, Horus 64 and Horus 128 recordings of the same excerpt from ‘A Midsummer Night’s Dream’. My Auralic Vega DAC meant I couldn’t evaluate the Horus 256, though.

 

The Mendelssohn is a piece that’s relatively easy to reproduce well, with good space, feather-light orchestration and nothing above mezzo forte.

 

Playback equipment was my normal system for hi-res digital downloads: An Auralic Aries bit cruncher, fed data from a high-speed Samsung 1Tb SSD; Auralic Vega DAC; Goldmund Mimesis 9 power amp and Martin Logan CLSIIz electrostatic speakers. Cabling in digital domain was by CAT, and Siltech monocrystal silver cable in analogue domain.

 

Anyway, my expectation was something like this: 1) that the Horus and Grimm would basically offer a different flavour but would not be radically different in overall SQ and 2) that the difference between the Horus 64 and Horus 128 would be quite subtle.

 

I was wrong on both counts. This was what I actually heard:

 

1) The Grimm 64 sounded markedly superior (IMO) to the Horus 64 – better focus, air, stage depth and timbral accuracy. Instruments sounded like the real thing.

2) The Horus 64 sounded more smeared in focus, less transparent, flatter in perspective and less musically accurate. I would hazard a guess that its time domain performance is/was not as good as that of the Grimm.

3) Horus 128 lifted the veil somewhat in focus and resolution, and also opened up some of the air and space. The difference between Horus 64 and 128 was more marked than I had expected.

4) However, I still preferred the Grimm overall. I’m not a Grimm groupie, but I have to say it impressed me mightily in these tests – mostly for musical reasons. Live music is my reference, and it simply sounded more to my ears like that!

 

And then Ted and Tom had to go and add two further tracks– respectively, the Horus DSD 256 downsampled to DSD 64 with Signalyst, and then the Signalyst-processed Grimm 64, with 0.34dB level tweak.

 

And here is where things started to get trickier. It would have been pretty easy to keep listening until I heard something but – in truth - comparing the two new tracks with the straight Grimm 64 these three seem to me, with my ears and playback equipment, now to be very much in the same ball-park, sonically. Yes, there are differences, but these seem much smaller than the large discrepancies that were pretty obvious between the straight Grimm 64 and downsampled Horus 64 and 128 tracks.

 

If this is valid, it suggests that the artefacts consequent on downsampling in the Merging DSD Converter software are sonically significant, whereas the Signalyst appears more transparent.

 

I would be happy musically with the straight Grimm or Signalyst tracks, less so it seems with those downsampled using Merging’s own software.

 

I hope this all makes sense!

 

I hope all is well with you.

Interesting to hear your take on these files .

I have only heard two versions so far the level corrected Grimm against Merging 128.

But I hear slightly more air from the Merging than the Grimm via Hugo and HD 800 headphones.

I will hopefully download and listen to the Merging 256 in Singapore soon.

The 128 took over an hour to download via slow wifi.

PS Have you got any new downloads to recommend?

Cheers Chris

[Jonalogic]

I hope all is well with you.

Interesting to hear your take on these files .

I have only heard two versions so far the level corrected Grimm against Merging 128.

But I hear slightly more air from the Merging than the Grimm via Hugo and HD 800 headphones.

I will hopefully download and listen to the Merging 256 in Singapore soon.

The 128 took over an hour to download via slow wifi.

PS Have you got any new downloads to recommend?

Cheers Chris

 

Chris Hi!

 

Nice to hear from you. Still bouncing around Asia, I see!

 

Well, I can't say the Mendelssohn results surprised me until I checked what tracks I'd been playing, and in what order. Yes, I randomised them, took notes and only checked at the end what was going on - all on the good scientific principle of blind testing...

 

I was surprised in the first round of tests at how the down-sampled Horus tracks sounded; it appears that extra processing step from its Native 256 files may have made the difference. When the Signalyst came in, it really seemed to equalise up the sonics of the Horus and Grimm. OK, there are still differences, but not as marked, and not apparently musically destructive ones.

 

I see many folk in the first round of tests preferred the Horus. How much of all of this may be due to equipment sweet spots and personal preference of listener ears is an open question...

 

To answer your last Q, frankly, not too much in the download arena has rung my chimes lately, But I have been doing some other things that are best discussed off-thread, perhaps, until I figure out whether they can be discussed in CA!

 

Cheers

 

Jon

[Jonalogic]

Jon, welcome to CA!! And thanks for the post. Yes, we are all becoming quite impressed with Miska's sampling routines. HQplayer is becoming a favorite, as I used it in my soon-to-be-finished exaSound e12 review, and now will use it to hopefully impress Jared with my surround setup (exaSound e28).

 

Ted Hi.

 

Many thanks for this. Happy to contribute.

 

Great test!

 

Cheers

 

Jon

[Jonalogic]

I hope all is well with you.

Interesting to hear your take on these files .

I have only heard two versions so far the level corrected Grimm against Merging 128.

But I hear slightly more air from the Merging than the Grimm via Hugo and HD 800 headphones.

I will hopefully download and listen to the Merging 256 in Singapore soon.

The 128 took over an hour to download via slow wifi.

PS Have you got any new downloads to recommend?

Cheers Chris

 

Chris Hi

 

Sorry I was bit coy in my earlier response. I didn't realise that you can safely mention SACD ripping in this fine Forum without an angry lynch mob with flaming torches appearing on your doorstep. Actually, I have been busy ripping a goodly selection of my SACD collection - for personal use only, naturally. The results have been startling and somewhat ear-opening.

 

But this might be better discussed - if you're interested - in the thread on CA devoted to SACD ripping.

 

Cheers

 

Jon

[filhodoassis]

My DSD playing system is on the cheap side: Foobar + SACD plugin DoP to GeekOut 450 (Sabre Chip) to a pair of nearfield monitors Adam Audio A5x + Sub7 subwoofer. Can play only DSD64 and DSD128.

In my system and to my ears, the best is Horus - Merging down to DSD128. Followed close by Grimm DSD64 original, and Horus - Sygnalist down to DSD64. These last two are near equivalent.

I wonder if another possible file, Horus - Sygnalist down to DSD128, woud be the better one in DSD64 and DSD128 camp, winning from Horus - Merging DSD128.

Many thanks for making this test possible, it is very informative.

[tailspn]

Good idea! I'll convert the Horus 256fs file using Sygnalist to 128fs, and upload it this evening. Since Jared is traveling to the AXPONA Show in Chicago, it will take a few days for the n/a to be replaced with the new Track 7 description on nativedsd.

[reverendo]

it's a pity that I can't make these tests, since my DAC doesn't do DSD, but I have to say that this is one of the best ideas that I've heard in a long time: significantly shortening the distance between consumer/enthusiasts/audiophiles and producers and recording engineers. simply fantastic!

[chrille]

Chris Hi!

 

Nice to hear from you. Still bouncing around Asia, I see!

 

Well, I can't say the Mendelssohn results surprised me until I checked what tracks I'd been playing, and in what order. Yes, I randomised them, took notes and only checked at the end what was going on - all on the good scientific principle of blind testing...

 

I was surprised in the first round of tests at how the down-sampled Horus tracks sounded; it appears that extra processing step from its Native 256 files may have made the difference. When the Signalyst came in, it really seemed to equalise up the sonics of the Horus and Grimm. OK, there are still differences, but not as marked, and not apparently musically destructive ones.

 

I see many folk in the first round of tests preferred the Horus. How much of all of this may be due to equipment sweet spots and personal preference of listener ears is an open question...

 

To answer your last Q, frankly, not too much in the download arena has rung my chimes lately, But I have been doing some other things that are best discussed off-thread, perhaps, until I figure out whether they can be discussed in CA!

 

Cheers

 

Jon

 

Yes I am indeed still bouncing around Asia. In fact on my way to KL for two live concerts this weekend and the week after at least one,maybe two in Singapore.

Nothing beats re-calibrating your ears with live music .

First Tchaikovsky's 1st piano concerto and Brahms's 4th with the MPO in KL .And then Holst's The Planets in Singapore coupled with a work I have maybe never heard live before, Strauss's Zarathustra.

Possibly at the Proms in the 70s or 80s when I heard basically the whole standard repertoire live every summer for many years.

Regarding this interesting comparison I have to admit my approach, partly for practical reasons of slow download connections, has not been very scientific.

I just wanted to hear the native Grimm against the Horus 128.

Maybe my conclusions so far,and only via headphones, are more than a bit expectation based.

I did hear a difference beween Horus 64 and native 128 on the comparison tracks from A far Cry so my expectations were that Horus at 64 would not be any better than Grimm at the same rate.

And with Heiliger Gedankensang from Beethoven's string quartet opus 132, sounding a bit airer native 128 rather compared to 64 I had the same expectations this time around.

It will be very interesting to hear both the native 256 and also the Signalyst converted 128 Tom has obviously uploaded now for further comparison. "The plot thickens".

I know that in theory as Ted writes in his article there shouldn't really be any audible differences between DSD 64 and 128 or even 256. But to my ears and obviously yours and many others here too,there are audible differences already between 64 and 128 in the first round.

 

Regarding downloads I have downloaded some of the old EMI Callas/ Karajan stuff you recommended,and like most of it a lot. Mainly for musical reasons.

But I have to admit that some of those old Blumlein Philharmonia Sibelius recordings are still pretty great sounding too.

I am interested to hear what path you have choosen for ripping SACDs.

Cheers Chris

[chrille]

Wouldn't one thing to keep in mind regarding the native Grimm be that we are now getting to hear the raw file without any editing or post processing involved? Under absolutely ideal conditions so to say. But once editing has been done for release that may also influence things, not necessarily in the favour of Grimm?

One of the known limitations and Achilles heels of with DSD 64 is the inevitable artefact high frequency noise, and it obviously adds up every time you edit if I have understood things correctly.

One of the differences I sometimes hear between DSD 64 recorded and hi res PCM is that with DSD 64 there can sometimes be a very high frequency,but still audible hiss/noise that I generally don't hear from say similarly recorded 24/176.4 or 24/192 or DXD recorded material.

And if there is noise it sounds a bit different from DSD noise.

Wouldn't the difference between noise starting at just above 20khz and 40khz or 80khz be reason enough to record with higher sampling rates whenever possible?

Or why not go all the way to 512 DSD?

And send the noise artefacts into space?

Just wondering.

[Jonalogic]

Yes I am indeed still bouncing around Asia. In fact on my way to KL for two live concerts this weekend and the week after at least one,maybe two in Singapore.

Nothing beats re-calibrating your ears with live music .

....

Possibly at the Proms in the 70s or 80s when I heard basically the whole standard repertoire live every summer for many years.

......

I did hear a difference beween Horus 64 and native 128 on the comparison tracks from A far Cry so my expectations were that Horus at 64 would not be any better than Grimm at the same rate.

And with Heiliger Gedankensang from Beethoven's string quartet opus 132, sounding a bit airer native 128 rather compared to 64 I had the same expectations this time around.

It will be very interesting to hear both the native 256 and also the Signalyst converted 128 Tom has obviously uploaded now for further comparison. "The plot thickens".

....

 

Regarding downloads I have downloaded some of the old EMI Callas/ Karajan stuff you recommended,and like most of it a lot. Mainly for musical reasons.

But I have to admit that some of those old Blumlein Philharmonia Sibelius recordings are still pretty great sounding too.

I am interested to hear what path you have choosen for ripping SACDs.

Cheers Chris

 

 

 

Chris Hi

 

I bet we must have run into each other without knowing it... the late 70s and 80s was the time when I too listened to just about everything live, either in the Proms or RFH in London. Then I buggered off to Hong Kong, but that's a whole other story.

 

We all know about the characteristic differences between PCM and DSD-flavours of playback. DSD 64 can get a bit thick and warm to my ears (almost like intermod?), but seems to become more transparent, focussed, tonally neutral and transparent as you go to 128. 256 I know nothing about - yet - as my Vega can't do it. But my gut feel is that DSD and PCM will converge in SQ as sampling/data rates increase further. I bet that DSD256 and DXD might sound sound pretty similar.... But of course I might be totally wrong on this. I often am.

 

To answer your final Q, I am using a hacked and reversioned PS3 CECHO3, using Ted's inimitable ripping guide. Damn me if the rips don't sound way better than my original SACDs. That stands to reason, as all the spinning disc rubbish, laser reading, jitter, error correction and other garbage is gone. Plus my Vega DAC is way better than that even in my trusty Esoteric X-O3SE SACD player.

 

But the real shocker is how much better the rips sound than some commercial-domain downloads I could mention - but won't.

 

Glad you liked the Callas recordings. Before she slimmed down too much and her wobble got endemic. Give me fat opera singers every day!

 

Cheers

 

Jon

[tailspn]

Signalyst Horus 256fs converted to 128fs now available as Track 7 Mendelssohn - BFO Horus 128fs via Signalist

[craighartley]

Signalyst Horus 256fs converted to 128fs now available as Track 7 Mendelssohn - BFO Horus 128fs via Signalist

 

Sorry if this is a dumb question, but what exactly is the Signalyst software in this context? I know HQPlayer does on the fly 'conversion', but how do you produce the same result offline?

[Sevenfeet]

First post.

 

Indeed, the hardest part of this test is actually playing a DSD256 file. I have an Oppo HA-2 portable headphone DAC (my first entry into Hi Res audio) and while the unit will certainly play DSD256 as part of its specs, no software I'm using now on my Mac will play it...Audirvana Plus converts to 24 bit/384 khz and Real Music and Amarra complete choke on it trying to play DSDs at all. And JRiver fails too, being able to play the DSD64 and DSD128 files but not the DSD256. On my iPhone, I can use the Onkyo player to listen to DSDs but it doesn't know how to handle the DSD256 either via bitstream.

 

The selection is very pretty to listen to. But I'm still trying to find a way to listen to this properly for a fair test.

[bmoura]

First post.

 

Indeed, the hardest part of this test is actually playing a DSD256 file. I have an Oppo HA-2 portable headphone DAC (my first entry into Hi Res audio) and while the unit will certainly play DSD256 as part of its specs, no software I'm using now on my Mac will play it...Audirvana Plus converts to 24 bit/384 khz and Real Music and Amarra complete choke on it trying to play DSDs at all. And JRiver fails too, being able to play the DSD64 and DSD128 files but not the DSD256. On my iPhone, I can use the Onkyo player to listen to DSDs but it doesn't know how to handle the DSD256 either via bitstream.

 

The selection is very pretty to listen to. But I'm still trying to find a way to listen to this properly for a fair test.

 

DSD 256 on the Mac may require a driver from the maker of your DAC (Oppo). For example, exaSound has ASIO drivers for their DACs that specifically support Stereo and Multichannel DSD 256 on the Mac or Windows.

 

DSD 256 Available on Mac for the First Time > exaSound Audio Design

[DaQi]

Wow, what a treat to be able to listen to these files and what a comparison. I was surprised by the results.

 

I downloaded all the 7 tracks available and played them in a random order. Actually, I thought I knew the order but when I sat down at my computer again they were not in the order I thought. So here are my results and thoughts...

 

Track 1: Horus @256 - Sounds muffled and boomy. Didn't like this one much.

Track 2: Horus - Sygnalyst @128 - This one sounded okay and much better than the first track

Track 3: Horus - Sygnalyst @64 - lots of static on this one almost sounded like a dirty record. It was so bad I almost got up to fiddle with things as I thought something was wrong with the system.

Track 4: Horus @128 - the static was gone so my system wasn't bad. This one seemed much more open than track 2 the dynamic range didn't sound problematic at all and the individual instruments sounded much clearer.

Track 5: Grim - Sygnalyst @64 - there was that static back again. Not as bad as in track 3 but still very annoying.

Track 6: Horus @64 - Hard to tell the difference between this and track 4. The dynamic range seemed slightly more tame than in track 4. I think I liked this one the best although track 4 and track 7 were close seconds.

Track 7: Grim @64 - Hard to tell the difference between this and track 6 but it sounded slightly less dynamic and a but muffled.

 

I was playing this from a fanless Intel NUC i5 using JRiver to a ifi Micro DSD then through my Rotel processor in bypass mode to my Rotel power amplifier and through to a pair of Linn Index speakers. I didn't touch any of the settings throughout the listening and left the processor at the same volume level throughout.

 

For my money the Horus 64 would be the one I would buy. I was surprised that the 256 didn't sound better and other than the tracks with nasty static on them it sounded the worst of the lot. The Horus at 128 was also quite nice as well. I think it might have been the clearest of them all but the Horus at 64 seemed to have the most presence and feeling in it.

 

Thanks again for the opportunity to get these files in their unadulterated raw form for a real apples to apples comparison.

[Miska]

Sorry if this is a dumb question, but what exactly is the Signalyst software in this context? I know HQPlayer does on the fly 'conversion', but how do you produce the same result offline?

 

I have provided an offline conversion tool for him that can do the rate conversions and level adjustments...

[Le Concombre Masqué]

TEAC 501(rest of the system is much more expensive but older too)/HQP were used (direct and not direct mode)

 

offering Horus generated files down converted at 64 would be a very very bad idea, imo ; that's by far the file I liked the less

 

Grimm's, especially HQP up sampled @128, hooked me up by the depth of the soundstage, its gentle welcoming sound that is inviting ; I got involved and listened

 

Horus @ 128 informs much more about the instruments : with the Grimm you know you listen to nicely recorded, i.e., string instruments and that's fine, nothing to complain ; with Horus 128 you might be able to count them and describe bowed instruments made out of wood, would you come from planet Mars and haven't seen or heard the description of one. Can't claim I much prefer it over Grimm's, though, different kind of beauties. too bad I can't play native Horus ; I bet I'd love it

 

 

HQP up sampling of 64 Horus to 128 does not at all sound like Horus 128 generated from 256 via Pyramix. direct vs not direct does not alter hierarchy nor deep nature of the files

[B&WQuad]

First post.

 

Indeed, the hardest part of this test is actually playing a DSD256 file. I have an Oppo HA-2 portable headphone DAC (my first entry into Hi Res audio) and while the unit will certainly play DSD256 as part of its specs, no software I'm using now on my Mac will play it...Audirvana Plus converts to 24 bit/384 khz and Real Music and Amarra complete choke on it trying to play DSDs at all. And JRiver fails too, being able to play the DSD64 and DSD128 files but not the DSD256. On my iPhone, I can use the Onkyo player to listen to DSDs but it doesn't know how to handle the DSD256 either via bitstream.

 

The selection is very pretty to listen to. But I'm still trying to find a way to listen to this properly for a fair test.

 

I have the same issue. Any suggestions for DSD256 support on a MAC. I know that Exasound has prop drivers for MAC. I am not aware of other manufacturer

[bmoura]

I have provided an offline conversion tool for him that can do the rate conversions and level adjustments...

 

Providing an offline upsampling and conversion app like that, to use with the HQ Player filters, would be quite a popular product for consumers as well.

 

Any plans to bring that to market?

[craighartley]

Providing an offline upsampling and conversion app like that, to use with the HQ Player filters, would be quite a popular product for consumers as well.

 

Any plans to bring that to market?

 

+1

I was about to thank Miska for his answer to my post and ask the same question.

[EESS]

The Grimm file is .3 db quieter than the Horus. Just enough to make sure the Horus files sound better.

 

Such a simple detail to get right.

 

The Grimm clock also is around 5 ppm slower though this should be of no audible consequence.

 

 

 

I converted the Horus 128 and the Grimm 64 to PCM. I cannot confirm a difference in loudness. But the frequency spectrum is considerably different. The Horus produces a slight decrease in volume up to 45 kHZ whereas the Grimm volume increase beginn 20 kHz and reaches the volume of 3kHz frequency at 40 kHz. An amazing difference.